The “pool” is mentioned in v18, v17, v16, v15, v14 documentation, under
- Language Reference > Web Server > Web server configuration and connection management
But there are not selectors that fit the description. I searched further, it is mentioned in v12, which is before the web server rewrite in v13, a tell-tale sign that it probably belongs to a different era.
Back then we had database parameters like
Minimum number of Web processes to maintain in non-contextual mode with 4D in local mode and 4D Server. By default, the value is 0.
Maximum number of Web processes to maintain in non-contextual mode with 4D in local mode and 4D Server. By default, the value is 10. In non-contextual mode, for the Web server to be reactive, 4D delays the Web processes for 5 seconds and reuses them to execute any possible future HTTP queries. In terms of performance, this is actually more advantageous than creating a new process for each query. Once a Web process is reused, it is delayed again for 5 seconds. When the maximum number of Web processes has been reached, the web process is then aborted. If no query has been attributed to a Web process within the 5 second delay, the process is aborted, except if the minimum number of Web processes has been reached (in which case the process is delayed again). These parameters allow you to adjust how your Web server operates in relation to the number of requests and the memory available as well as other parameters.
One cannot but notice that the description of how these contribute to performance is pretty out of date.
So I would say that this is a documentation error.